Wavelet analysis for velocity characterization in mobile networks

This study deals with the issue of discriminating between different classes of mobile velocity (incar, pedestrian, immobile), based on the multiresolution analysis of the measured radio signals. We make use of the decision theory in order to build an automatic classifier of radio signals. This begins by a feature extraction based on a wavelet decomposition, followed by a statistical classifier to decide about the class of the radio signal. Our method is applied on real measurement signals and the results prove that a simple wavelet-based mobility characterization is possible using the available radio measurements, even though they are averaged and under-sampled